Long-term culture of human brain organoids
• We recommend pilot experiments to find the optimal number of starter cells for embryoid body formation, testing cell seeding densities between 2,000 and 3,000 cells per well. The optimal starter cell number may vary depending on the proliferation rates of different PSC lines.
• For embryoid body formation, use only cells that do not display spontaneous differentiation.
• Troubleshooting using different PSC lines may be necessary if the specific PSC line has not been used to generate brain organoids, as there are differences in efficiency of neuroectodermal differentiation across different PSC lines.
• In order to reduce variability between organoids cultured long-term, screen organoids after 5 days in neural induction conditions before embedding in Matrigel. Select for embedding only organoids with visible formation of neuroepithelium. Neuroepithelial formation is demonstrated by the presence of translucent tissue around the edges of the organoid.
• The incubation time in papain solution required for dissociation may vary with organoid age and between organoids generated from different cell lines.
Electrophysiological Recordings
• How do I know that the silicon probe is submerged in the chamber?
If you are using an audio amp, you will hear a distinctive change in sound when the probe hits the solution surface and the circuit closes. You can also monitor the level of channel noise using the SmartBox software, or do a quick impedance measurement. A probe that is not in contact with the aCSF will have an open circuit and a very large impedance.
• Problems Anchoring the Organoid
The majority of organoids will fall through the chute and come to rest on the anchor with little additional assistance. Some organoids may roll off the anchor. These troublesome organoids can be carefully manipulated onto the anchor using two fine paintbrushes.
• Chamber temperature keeps rising
Check that the heat platform / feedback thermistor is in-place. If the thermistor has poor contact with the aluminum heating plate or fits loosely, it can respond poorly and lead to excessive heating of the recording chamber. Thermal transfer can be improved by adding a drop of mineral oil or vacuum grease to the thermistor and its mounting hole. Note that high temperature can be detrimental to the tissue.
• Chamber temperature is low
Check that the bath thermistor is submerged in the recording chamber. Note that this thermistor only monitors temperature.
• Heating stage is not warming up
Check that the heating element (power film resistor) is functional, and that it has good thermal contact with the heating stage. Thermal transfer can be improved by placing a heat transfer pad between the resistor and the stage.
• What is the maximum temperature that can be used to drive the heating stage?
The heat stage thermistor has a maximum temperature rating of 60°C, and this should not be exceeded by the power film resistor / heating element (the power film resistor has a maximum temperature of 150°C). If the setup is functioning correctly, there should be no need to supply temperatures beyond the thermistor’s rating. Check that all component of the homeothermic feedback system are functional, and that there is good thermal contact at all sites.
• Chamber pH drifts
Check that the carbogen bubbling is adequate for bicarbonate buffering.
Check that the chamber lid is in place. The lid will help to minimize evaporation and maintain salt concentrations.
Tip: With long recordings a drift in the pH with the described setup is unavoidable. Long recording will benefit from adapting the recordings chamber to a perfusion system. Intermediate recording lengths might benefit from switching to a HEPES-based buffer.
• The wide-band signal is dominated by 50-60 Hz line noise
If the rig is optimally configured, 50-60 Hz noise should be negligible / invisible by eye in the time-domain. If you experience prominent line noise in the wide-band signal, the rig will require some additional optimization:
clean up wiring to remove ground loops
check that the system is grounded correctly
check that the AgCl ground / ref is in good condition
check the quality of soldering
check that aCSF is not overflowing from the recording chambers
• Intermittent low frequency noise in the wide-band signal, synchronized across probe sites
Instability in the baseline signal could be related to the carbogen flow. Bursting of large bubbles can produce surface ripples that are capable of producing recording artifacts. To fix, adjust the flow rate and the positioning of the gas line.
• What is the Expected Noise?
Using the system described here, for new high density silicon probes, the typical signal root-mean-square (rms) is 3-4 μV. Conservatively, with on-line peak thresholding, it is possible to isolate spikes >24 μV.
• Channel noise is larger than expected
With repeated reuse of silicon probes there is a build up of detritus, paralleled by increase in probe impedance and the rms. This will negatively impact the signal to noise ratio. We recommend replacing the probe when the majority of channels have >8 μVrms.
• How do I extend the life of the silicon probes?
The lifespan of probes can be extended by rinsing in dd.H2O immediately after use, and then storing the silicon shaft in a mild protease or detergent (for example, a contact lens cleaning solution suitable for silicon lenses). Use of alcohol is NOT recommended, as it will cause tissue debris to dehydrate and adhere strongly to the probe. Old probes can be reconditioned by chemical stripping and re-coating with PEDOT.
• Channel Impedance is low (<0.1 MΩ, measured at 1000 Hz in aCSF)
This is indicative of a short-circuit. Channels with low impedance should be discarded from analysis.
Tip: Disabling acquisition of data from defective channels might produce changes to file mapping. If memory is not an issue it can be simpler to acquire all channels and remove defective channels during analysis.
• Channel impedance is high (>1 MΩ, measured at 1000 Hz in aCSF)
High impedance channels should be discarded from analysis. We have often observed spike-like events that are isolated to high-impedance channels; these events will not appear on neighboring sites with an expected impedance. As these events / artifacts often occur synchronously at multiple high-impedance probe sites, and do so irrespective of their spatial geometry, it is not believed that they are spikes.
Tip: Disabling acquisition of data from defective channels might produce changes to file mapping. If memory is not an issue it can be simpler to acquire all channels and remove defective channels during analysis.
• During recordings, spikes drift across probe sites
Probe implantation can lead to some small compression of the surrounding tissue. Compressed tissue will settle back / rebound over time, and can cause spikes to exhibit a time-related drift across probe space. If this movement is large it can hamper the clustering of spikes during analysis. Drift is particularly troublesome when there are multiple neuronal units whose spikes overlap in probe space. The greater the overlap, and the more numerous the units, the greater the problem in spike sorting. If this is a problem, position the probe and then let the tissue settle/recover for 10-15 minutes before beginning the experiment proper.
• The spikes signal to noise ratio is small
Check the quality of the probe.
Try repositioning the probe (recorded spikes are larger when the probe is proximal to the cell soma6).
Note: Some contributing factors are intrinsic to the neuron being recorded from6, and outside of the experimentalist’s control. It should also be noted that key features of a neuron change during maturation (for example: soma size, dendritic morphology, ion channel expression and distribution), and will likely produce a recording bias towards mature neurons.
Tip: All errors caused by the assignment of spikes to the wrong neuronal unit, or the dropping of spikes that fall below a detection threshold, propagate detrimentally down the analysis stream. Therefore, analysis should only be done on well-isolated spike trains.
• How deep should I advance the probe tip?
If the aim is to record from this superficial population of neurons, the probe should be advanced by a distance equal to the vertical spread of the probe sites. The high-density silicon probe described in this protocol are composed of 64 9x9 μm probe sites, organized in 2 columns of 32 rows, with a 12 μm pitch2; to record from neurons at the organoid’s surface, these 2x32 probes should be advanced 380 μm.
• Is the composition of aCSF important?
The composition of aCSF, and the recording temperature, was selected to be close to physiological 7,8. Modifications to aCSF composition might change the level / structure of spontaneous network activity.
• How do I perform acute pharmacological studies?
To avoid the effects caused by the dilution of aCSF it is recommended that negligible volumes of stock solution are added to a known volume of aCSF (15 mL in this protocol). TTX, applied to the external chamber, typically attenuates neuronal spike rate within 2-5 minutes (this is based on recordings made from probes implanted <380 μm below the organoids surface).
Tip: Internal control experiments should always be performed. For example, add an equivalent volume of the vehicle without drug.
Tip: Consider adapting the described setup to a perfusion-based recording chamber; this will enable more elegant pharmacological protocols.
• What is the best way to clean the recording chamber?
We recommend rinsing immediately after use with dd.H2O. Repeat the rinse to remove salts and tissue debris.
Note: Calcium deposits will form if the rinsing procedure is insufficient.
Tip: The process of removing solution from the recording chamber can be simplified by setting up a vacuum aspirator. For precision aspiration, we recommend using a 200 μL micropipette tip. To increase the speed of aspiration trim back the end of the micropipette tip.